In the attempt to conceive a process by which Evolution may have come about, the first phenomenon to be recognized and accounted for is specific difference. With that recognition the outline of the problem is defined. The second prerogative fact is adaptation. Forms of life are on the whole divided into species, and these species on the whole are adapted and fit the places in which they live. To many students of Evolution, adaptation has proved so much more interesting and impressive than specific diversity that they have preferred it to the first place in their considerations.

Whether this is, as I believe, an inversion of the logical order or not, there is one most serious practical objection to such preference, that whereas specific diversity is a subject which can be investigated both by the study of variation and by the analytical apparatus which modern genetic science has developed, we have no very effectual means of directly attacking the problems of Adaptation.

The absence of any definite progress in genetics in the last century was in great measure due to the exclusive prominence given to the problem of Adaptation. Almost all debates on heredity centered in that part of the subject. No one disputes that the adaptation of organisms to their surroundings is one of the great problems of nature, but it is not the primary problem of descent. Moreover, until the normal and undisturbed course of descent under uniform conditions is ascertained with some exactness, it is useless to attempt a survey of the consequences of external interference; nor as a rule can it be even possible to decide with much confidence whether such interferences have or have not definite consequences. Those, for example, who debated with enthusiasm whether acquired characters are or are not transmitted were constantly engaged in discussing occurrences which we now know to be ordinary features of descent under uniform conditions, and the origin of variations which were certainly not caused directly by circumstances at all. In the absence of any factorial analysis, or of any conception of what factorial composition means and implies, no one knew what varieties might be expected from given parents. The appearance of any recessive variety was claimed as a consequence of some treatment which might have been applied to the parents. There was no possible standard of evidence or means of controlling it, and thus the discussion was singularly unfruitful. Before we can tell how the course of descent has departed from the normal, we must know what the normal would have been if we had let alone. We are still far from having such knowledge in adequate measure, but it does now exist in some degree, and we are steadily approaching a position from which we shall be able to form fairly sound estimates of the true significance of evidence for or against the proposition that environmental treatment can produce positive disturbances in the physiological course of descent.

Thus described, the field for consideration is very wide. Though the effects of changed conditions were especially studied in the hope of solving the problem of adaptation by direct observation, that, as all are now agreed, is but a part of a more general question. We must ask not only do changed conditions produce an adaptative response on the part of the offspring, but whether they produce any response on the part of the offspring at all. It is not in doubt that by violent means, such as starvation or poisoning of the reproductive cells, effects of a kind, stunting and deformity for instance, can be made evident, just as similar effects may follow similar treatment during embryonic or larval life. Apart from interferences of this class, are there any that may be reasonably invoked as modifying the course of inheritance?

No epitome of the older evidence for the inheritance of adaptative changes is here required. That has often been collected, especially by Weismann, who exposed its weaknesses so thoroughly as to carry conviction to most minds, and showed that whether the phenomenon occurs or not, no one can yet prove that it does. Belief in these transmissions, after being almost universally held, was with singular unanimity abandoned. This change in opinion, though doing credit to the faith of the scientific community in evidential reasoning, is the more remarkable inasmuch as the strength of the idea was not derived from the minute amounts of supposed facts now demolished. On the contrary, it was really an instinctive deduction from a wide superficial acquaintance with the properties of animals and plants. They can accommodate themselves to circumstances. They do make responses sometimes marvellously appropriate to demands for which they can scarcely have been prepared. What more natural than to suppose that the permanent adaptations have been achieved by inherited summation of such responses? No one had actually been driven to believe in the inheritance of adaptative changes because bitches which had been docked had been known to give birth to tailless puppies, or because certain wheat in Norway was alleged to have become acclimatized in a few generations. Evidence of this kind was collected and produced rather as an ornamental appendix to a proposition already accepted, and held to be plainly demonstrated by the facts of nature. Looked at indeed in that preliminary and uncritical way, the case is simply overwhelming. Those who desire to see how strong it is should turn to Samuel Butler's Life and Habit, and even if in reading they reiterate to themselves that no experimental evidence exists in support of the propositions advanced, the misgiving that none the less they may be true is likely to remain. Making every deduction for the fact that the wonders of adaptation have been grossly exaggerated, and that marvels of fitness and correspondence between means and ends have grown out of mere anthropomorphic speculations, there is much more left to be accounted for than can at all comfortably be accepted as the product of happy accidents. So oppressive are these difficulties that we can scarcely blame those who imagine that the study of heredity is primarily directed to the problem of the transmission of acquired characters, a preconception still almost universal among the laity.

But since the belief in transmission of acquired adaptations arose from preconception rather than from evidence, it is worth observing that, rightly considered, the probability should surely be the other way. For the adaptations relate to every variety of exigency. To supply themselves with food, to find it, to seize and digest it, to protect themselves from predatory enemies whether by offence or defence, to counter-balance the changes of temperature, or pressure, to provide for mechanical strains, to obtain immunity from poison and from invading organisms, to bring the sexual elements into contact, to ensure the distribution of the type; all these and many more are accomplished by organisms in a thousand most diverse and alternative methods. Those are the things that are hard to imagine as produced by any concatenation of natural events; but the suggestions that organisms had had from the beginning innate in them a power of modifying themselves, their organs and their instincts so as to meet these multifarious requirements does not materially differ from the more overt appeals to supernatural intervention.

The conception, originally introduced by Hering and independently by S. Butler, that adaptation is a consequence or product of accumulated memory was of late revived by Semon and has been received with some approval, especially by F. Darwin. I see nothing fantastic in the notion that memory may be unconsciously preserved with the same continuity that the protoplasmic basis of life possesses. That idea, though purely speculative and, as yet, incapable of proof or disproof contains nothing which our experience of matter or of life at all refutes. On the contrary, we probably do well to retain the suggestion as a clue that may some day be of service. But if adaptation is to be the product of these accumulated experiences, they must in some way be translated into terms of physiological and structural change, a process frankly inconceivable.

To attempt any representation of heredity as a product of memory is, moreover, to substitute the more obscure for the less. Both are now inscrutable; but while we may not unreasonably aspire to analyse heredity into simpler components by ordinary methods of research, the case of memory is altogether different. Memory is a mystery as deep as any that even psychology can propound. Philosophers might perhaps encourage themselves to attack the problem of the nature of memory by reflecting that after all the process may in some of its aspects be comparable with that of inheritance, but the student of genetics, as long as he can keep in close touch with a profitable basis of material fact, will scarcely be tempted to look for inspiration in psychical analogies.

For a summary of the recent evidence I may refer the reader to Semon's paper[1] where he will find a collection of these observations described from the standpoint of a convinced believer. At the outset one cannot help being struck by the fact that of the instances alleged, very few, even if authentic, show the transmission of acquired modifications which can in any sense be regarded as adaptative, and many are examples not so much of a transmission of characters produced in the parents as of variation induced in the offspring as a consequence of treatment to which the parents were submitted, the parents themselves remaining apparently unmodified. No one questions the great importance of evidence of this latter class as touching the problem of the causes of variation, but it is not obvious why it is introduced in support of the thesis that acquired characters are inherited.

It is most difficult to form a clear judgment of the value of the evidence as a whole. To doubt the validity of testimony put forward by reputable authors is to incur a charge of obstinacy or caprice; nevertheless in matters of this kind, where the alleged phenomena are, if genuine, of such exceptional significance, belief should only be extended to evidence after every possible source of doubt has been excluded. We believe such things when we must, but not before. At the very least we are entitled to require that confirmatory evidence should be forthcoming from independent witnesses. So far as I have seen, this requirement is satisfied in scarcely any of the examples that have been lately published, and until it is, judgment may reasonably be suspended.

In some cases, however, the facts are not doubtful. Standfuss, by subjecting pupae of Vanessa urticae to cold, produced the now well-known temperature-aberrations in which the dark pigment is greatly extended. He put together in a breeding-cage 32 males and 10 females showing this modification in various degrees. Two of these females died without leaving young. Seven produced exclusively normal offspring. From the eighth female 43 butterflies were bred, and of these there were four (all males) which to a greater or less extent exhibited the aberrational form.[2] The mother of this family was the most abnormal of the 10 females originally put in.

Fischer's experiment with Aretia caja was on similar lines. From pupae which had been frozen almost all the moths which emerged showed aberrational markings. A pair of these mated and produced 173 young which pupated. Those which emerged early were all normal, but of those which emerged late, 17 had in various degrees abnormal markings like those of the parents.[3] In neither of these examples is there any question as to the facts. Both observers have great experience and give full details of their work.

As regards Vanessa urticae, however, it must be recalled that Fischer himself showed that in Nymphalids somewhat similar aberrations could be produced both by heat and by cold, and even by centrifuging the pupae. Frl. von Linden produced a transitional form of the same aberration in V. urticae by the action of carbonic acid gas.[4] It is highly probable that the appearance is due to a morbid change, perhaps an arrest of development, which may be brought about by a great diversity of causes. In the experiments the cause probably was a diseased condition of the tissues of the mother herself. She had been subjected to freezing sufficiently severe to prevent the proper development of the pigments and some of the ovarian cells presumably suffered also. It will be observed that the only specimens which were affected were the offspring of the most abnormal female, and of them only four out of forty-three showed any change.

The same interpretation probably applies to the cases in Arctia caja. In this species the markings are well known to be liable to great variation. As Barrett says, even in nature individuals are rarely quite alike, and an immense number of strange forms occur in collections.[5] These are greatly sought after by some collectors, especially in England, where they fetch high prices at auctions, and it is notorious that most of them come from Lancashire and the West Riding of Yorkshire. It is commonly supposed that the breeders of that district subject them to abnormal conditions, and especially to unnatural feeding, but I know no clear evidence that this is true. From whatever cause it is certain that the natural pattern is, in some strains at all events, very easily disturbed.

The elaborate experiments of SchrÖder with Abraxas grossulariata are difficult to follow and are complicated by the fact that the series which was submitted to abnormal temperatures was derived from an abnormal original pair. From the evidence given it is not clear to me whether the temperature had a distinct effect. This insect, like Arctia caja, produces an immense number of variations (especially in the amount of the black pigment) and as most of these are, I believe, reared in domestication for sale, it is highly probable that the species is easily influenced by cultural conditions.

SchrÖder describes two other experiments which have been accepted by Semon and other supporters of the view that acquired characters are transmitted. In the first, Phratora vitellinae, a phytophagous beetle living on the undersides of leaves, was used. It naturally feeds on Salix fragilis, a species without a felt, or tomentum, on the underside of the leaves. Larvae were transferred to another willow (near S. viminalis) which has the undersides of the leaves felted. The larvae took readily to the new food, pushing the tomentum before them as they gnawed the leaves. They came to maturity and when they were about to lay their eggs they were given a free choice between S. fragilis and the tomentose species. The greater number of ovipositions, 219, took place on fragilis, and there were 127 on the tomentose bush, which we are told was six times as large as the fragilis. The larvae from fragilis were next put on the tomentose species and reared on it. When they became imagines they were similarly given their choice, with the result that there were 104 ovipositions on the tomentose species and only 83 on fragilis. In the next generations there were 48 ovipositions on the tomentose and 11 on fragilis. Finally the fourth generation made 15 ovipositions on the tomentose and none on fragilis.

The difficulty about such experiments is obviously that one has no assurance that the change of instinct, in so far as there is any, may not be a mere consequence of the captivity. It must, besides, be extremely difficult to arrange the experiment so that there is really an equal choice between the two bushes, when one stands beside the other. Przibram, in quoting this case, considers that as the tomentose bush was about six times as large as the fragilis, some indication of the relative attractiveness of the two may be obtained by dividing the ovipositions on the larger bush by six, but I imagine the matter must be much more complex.

SchrÖder's second example is not more convincing, in my opinion, though Semon regards it as one of the most important pieces of evidence. It concerns a leaf-rolling moth, Gracilaria stigmatella, the larva of which is said normally to make its house by bending over the tips of the sallow leaves on which it feeds. SchrÖder placed larvae on leaves from which the tips had been cut, and these larvae made their houses by rolling over the sides of the leaves. Their offspring were again fed on leaves without tips, and as before, they rolled in the leaf-margins either on one side or both. The offspring of this second generation were then fed on entire leaves. There were 19 houses made by these (?19) larvae, and of them 15 were normal, made by folding down the tips of the leaves, while 4 were abnormal, made by rolling in the leaf-margins. SchrÖder says that in nature he has only twice seen abnormal houses; but it is clearly essential not only that the frequency of such variability in nature should be thoroughly examined, but also that we should know whether when the species is bred in captivity these irregularities of behaviour do or do not occur when the larvae are fed on uninjured leaves.

The famous case of SchÜbeler's wheat is revived by Semon. The story will be familiar to most readers of the literature of the subject. Briefly it is that annuals, especially wheat and maize, raised from seed in Central Europe take more time in coming to maturity and ripening than similar plants raised in Norway, where the summer days are much longer. The received account is that he imported seed especially of maize and of wheat from Central Europe to Norway and found that in successive years the period of growth and ripening was increasingly reduced. After two generations seed of the accelerated wheat was sent back to Breslau where it was grown, and was found to ripen rather more slowly than in Norway, but much more quickly than the original stock had done. The facts recorded by SchÜbeler[6] are that he received seed from Eldena, which is on the Baltic near Greifswald. The variety is described as "100 tÄgiger Sommer Weizen," but no more exact record of its behaviour in Germany is given. This wheat, grown at Christiania in 1857, took 103 days to harvest. Its seed was again grown in Christiania in 1858, and took 93 days, and sown again in 1859 it took only 75 days, 28 days less than in the first year of cultivation in Norway. Seed of the 1858 crop was sent to Breslau, and grown there by Roedelius in 1859; it took 80 days. Evidently before such a record can be used as proving an inheritance of acquired characters numbers of particulars should be forthcoming. The view that Johannsen has taken is that the result was probably due to unconscious selection of the earlier individuals among a population consisting of many types of various compositions. Some effect may no doubt be ascribed to that cause, but I cannot think that alone it would account for the results. My impression is rather that they were produced by differences in the cultivation and especially in the seasons. Research of an elaborate character would be necessary in order to eliminate the various sources of error, and nothing of the kind has been done; nor does Semon allude to these difficulties in prominently adducing SchÜbeler's evidence. A difference of even three weeks in time of harvesting may easily be due to variation in the season. It would in any case be difficult to analyse the meteorological conditions, and to decide how much effect in postponing or accelerating the harvest might be due to cold days, to cloudy days, to wet weather, to fluctuations in average temperature, to hot days, and other such incidents occurring at the different periods of growth, even if they were specially watched while the experiments were in progress, and at this distance of time such analysis is practically impossible. Without careful simultaneous control-experiments this evidence is almost worthless. The director of the Meteorological Office[7] has, however, kindly sent me some details of the weather at Breslau from 1857 to 1860, and I notice that as a matter of fact July, 1859, was an exceptionally hot month, having an average of 2.67° C. above the mean for the twenty years 1848-1867. June in that year was slightly (0.31° C.) below the mean and May slightly above it (0.18° C.). August was also abnormally hot, 2.35° C. above the average. The Breslau wheat was sown on May 19 and harvested on August 6. There was a cold spell from May 11 to 14, which this wheat escaped, as it was sown on May 19. In the other years the cold spell came much later. These elements of the weather may possibly have done something to hurry the ripening in 1859. It unfortunate that we are not told how long similar wheat from Breslau seed took to ripen in that year.

As regards the Norway cultivations we have the average monthly temperatures recorded by SchÜbeler, though he does not discuss them in connection with this special problem. It is quite clear that 1857, in which the period was 103 days, was an exceptionally cold summer, especially as regards the months of June and July, but though there was, so far as the temperature records go, no great difference between 1858 and 1859, the year 1859, in which the period of ripening was the shortest, was somewhat colder in Norway than 1858. But we have the further difficulty that there were ten days difference in sowing, for in 1858 the sowing was made on May 14, and in 1859 on May 24. With all these possibilities uncontrolled, and indeed unconsidered, I am surprised that Semon should claim these experiments as one of the chief supports for his views.

SchÜbeler's other allegations respecting the influence of climate on plants grown in various places and especially at different elevations in Norway have been destructively criticised by Wille[8] to whose paper readers interested in the subject should refer.

Before the appearance of Wille's criticisms Wettstein[9] made a favourable reference to SchÜbeler's work, accepting his conclusion. He states also that he has himself made analogous experiments with flax, finding that the length of the period of development and a series of morphological characters show an adaptation to local conditions, and that on transference of seed to other conditions the previous effects are maintained. No details, however, are given, and I do not know if anything more on the subject has appeared since. The other examples cited by Wettstein, such as the observations of Cieslar on forest-trees and those of Jakowatz on gentians seem to me open to all the usual objections applicable to evidence of this kind. Such work, to be of any value for the purpose to which it is applied, must be preceded by a study of the normal heredity and of the variations of the species.

Most of the recent writers (Semon, Przibram, etc.) on the inheritance of acquired characters accept the story of Brown-SÉquard's guinea pigs, which are said to have inherited a liability to peculiar epileptiform attacks induced in their parents by various nervous lesions.

The question has been often debated and several observers have repeated the experiments with varying results, some failing to confirm Brown-SÉquard, others finding evidence which in various degrees supported his conclusions. Recently a new and especially valuable paper has been published by Mr. T. Graham Brown[10] which goes far towards settling this outstanding question. He states that "the Brown-SÉquard phenomenon is nothing more or less than a specific instance of the scratch-reflex," and it is due to a raised excitability of the mechanism of this reflex. This raised excitability is the character acquired as a consequence, for instance, of the removal of part of one great sciatic nerve. The nature of this raised excitability and its causation are discussed and elucidated, but this part of the work is not essential to the present consideration. Mr. Graham Brown in his summary of conclusions remarks that it is very difficult to see how this condition of raised excitability can be transmitted to the offspring, and this comment which might be made in reference to any of the alleged cases certainly applies with special cogency to the present example.

He then calls special attention to three observations:

1. That guinea pigs which had a "trophic" change in the foot, as a result of division of the great sciatic nerve, have repeatedly been seen to nibble the feet of other guinea pigs which had this change in the foot from the same causes.

2. That accidental injury to the toes may be followed by the Brown-SÉquard phenomenon in an otherwise normal animal.

3. That in several instances the young of guinea pigs which exhibited the phenomenon have been noticed to have one or more toes eaten off by the mother.

Brown-SÉquard noticed that almost all his animals in which the great sciatic was divided acquired the "epilepsy" and nibbled those parts of their feet in which sensation had been lost. Of the offspring of such animals he found that a very small proportion exhibited a malformation of the feet, and of these some showed the "epilepsy." The proportion which showed the "epilepsy" was one to two per cent. of the offspring.

Morgan[11] is quoted by Graham Brown as having suggested that the loss of toes in the offspring may have been due to mutilation by the mother, following his experience in a case in which the tails of mice in succeeding litters were thus devoured, and there can be little doubt that in this suggestion lies the clue to the explanation of the whole mystery. Graham Brown concludes that it may be supposed with every degree of probability that the "transmission" was due to injuries inflicted upon the young by their parents. With this conclusion most people will now be disposed to agree, and we may hope that we shall hear the last of this curious myth—to the elucidation of which a vast quantity of research has been devoted.

The series of experiments made by Kammerer with various Amphibia have attracted much attention and have been acclaimed by Semon and other believers in the transmission of acquired characters as giving proof of the truth of their views. With respect to these observations the chief comment to be made is that they are as yet unconfirmed. Many of the results that are described, it is scarcely necessary to say, will strike most readers as very improbable; but coming from a man of Dr. Kammerer's wide experience, and accepted as they are by Dr. Przibram, under whose auspices the work was done in the Biologische Vesuchsanstalt at Vienna, the published accounts are worthy of the most respectful attention.

The evidence relates chiefly to three distinct groups of occurrences:

1. Modification in Alytes obstetricans, the Midwife Toad, affecting both the structure and the mode of reproduction, induced by compulsory change of habits.

2. Modification in the mode of reproduction of Salamandra atra and maculosa induced by compulsory change of habits.

3. Modification in the colour of Salamandra maculosa induced by change in the colour of the soil on which the animals were kept.

1. I will take first the case of Alytes,[12] because it is the most definite example, and because it is the case which most readily admits of repetition and verification.

The habits of Alytes obstetricans are well known. The animals copulate on land. As the strings of eggs leave the female they are entangled by the hind legs of the male, and being adhesive they stick to him and undergo their development attached to his back and legs. The number of eggs varies from 18 to 86, a number much smaller than is usual in toads and frogs which lay their eggs in water. The eggs are large and full of yolk.

There are two breeding seasons, one about April and the other about September, and a winter hibernation. Not only animals brought in from outside, but their offspring reared in domestication maintain these normal habits in confinement, if the temperature does not exceed 17° C. (pp. 499 and 534).

If, however, the temperature be artificially raised and kept at 25-30° C., the males do not attach the eggs to themselves when spawning occurs on land but let them lie. The adhesion of the eggs is said to be hindered by the comparatively rapid drying of their surfaces.

More usually in the high temperatures the animals take to the water and copulate there. The eggs are ejected into the water, and as their gelatinous coverings immediately swell up, they do not stick to the males.

The offspring thus derived from the parents subjected to heat for one breeding-period only, whether they were laid in water or on land, did not show departures from the normal type.

Kammerer states next, however, that in subsequent breeding-periods the same parents frequently take to the water to breed, though they have become quite accustomed to the heated chamber; and furthermore that if such animals, having thus lost their instinct to brood their young, be transferred to ordinary temperatures they do not readily reassume their normal habits, but for several breeding seasons—at least four—will take to the water. These parents lay from 90 to 115 eggs, which are small and contain little yolk, and the larvae, on hatching, breathe with their embryonic gills until they are absorbed instead of being broken off as normally.

The offspring thus abnormally developed when they mature are said never to brood their eggs. If they are derived from the earlier spawnings of their parents, before, that is to say, the parents had been submitted to the changed conditions long enough to transmit their effects, they lay on land; but if they are derived from the later spawnings, they lay in the water. These changes of habit are manifested without the continued application of the abnormal experimental conditions, and, as I understand the account, in normal conditions of temperature.

If the abnormal experimental conditions are continued, the toads always lay in water, and their eggs become progressively smaller and more numerous. The larvae in the fourth generation acquire three pairs of gills instead of one pair, and are in other respects also different from the normal form.

Respecting the Alytes bred in this way Kammerer makes the very striking statement that the males in the third generation (p. 535) have roughened swellings on their thumbs and that in the fourth generation (pp. 516 and 535) these swellings develop black pigment. Together with the appearance of this secondary sexual character there is hypertrophy of the muscles of the fore-arm. To my mind this is the critical observation. If it can be substantiated it would go far towards proving Kammerer's case. Alytes, among toads and frogs, is peculiar in that the males do not develop these lumps in the breeding season, and the fact may no doubt be taken to be correlated with the breeding habits, copulation occurring on land and not in water as is usual with Batrachians. It is to be expressly noticed that these lumps on the thumbs or arms of male toads and frogs are not merely pigmented swellings, but are pads bearing numerous minute horny black spines, which are used in holding the females in the water. The figures which Kammerer gives (Taf. XVI, figs. 26 and 26a) are quite inadequate, and as they merely indicate a dark patch on the thumbs it is not possible to form any opinion as to the nature of the structure they represent.

The systematists who have made a special study of Batrachia appear to be agreed that Alytes in nature does not have these structures; and when individuals possessing them can be produced for inspection it will, I think be time to examine the evidence for the inheritance of acquired characters more seriously. I wrote to Dr. Kammerer in July, 1910, asking him for the loan of such a specimen[13] and on visiting the Biologische Versuchsanstalt in September of the same year I made the same request, but hitherto none has been produced. In matters of this kind much generally depends on interpretations made at the time of observation; here, however, is an example which could readily be attested by preserved material. I notice with some surprise that in a later publication on the same subject no reference to the development of these structures is made (see below).

The statements here given represent but a small part of Kammerer's papers on the subject. He gives much further information as to the course of the experiments, especially in regard to the fate of the eggs laid on land and the aberrations induced in them by treatment. The ramifications of the experiments are, however, very difficult to follow, and as I am not sure that I have always understood them I must refer the reader to the original.

More recently Kammerer has published[14] a most curious account of experiments in crossing his modified and abnormal Alytes, derived from the water-eggs, with normal individuals.

In the first case the cross was made between a normal female and an abnormal male. The offspring were normal in their habits. In the next generation bred from these almost exactly a quarter showed the abnormal instinct.

The reciprocal cross was made between an abnormal female and a normal male. In this case the offspring were abnormal in their behaviour; but the second generation bred from them showed three quarters abnormal and one quarter normal.

Certain details as to numbers and sexes of the various families bred in the course of this amazing experiment are given in a subsequent publication.[15] This later paper goes somewhat fully into the question of the difference in behaviour between the normal and modified individuals, describing the ways in which the males and females possessing the acquired character could be recognised from the males and females which were normal, but in this account I find no reference to the development of the "Brunftschwielen"—the horny pads on the hands of the males. As these structures would be of special value in such a diagnosis the omission of any allusion to them calls for explanation. Kammerer claims the evidence as proof of Mendelian segregation in regard to an acquired character, the first example recorded. Pending a repetition of the experiments there is no more to be said.

2. The Mode of Reproduction of Salamandra atra and maculosa.[16] mdash;Salamandra maculosa, the common lowland form, with yellow bands or spots, deposits its young in water, generally as gill-bearing tadpoles, with a wide, swimming tail, though occasionally they are born still enclosed in the egg-capsule out of which they soon hatch. Spawning extends over a considerable period, often many weeks, and during the season one female may bear more than 50 young.

S. atra, the black Alpine form, produces its young on land. They are born without gills, ready to breathe air, and with the rounded tail of the adult. These differences may, as Kammerer says, naturally be regarded as adaptations to the Alpine conditions. Moreover, the female bears only two young in a season, and this reduction in the number must be taken to be a consequence or condition of viviparity. There are many eggs in the ovary, but all except the two which are destined to develop degenerate and form a yolk-material on which these two survivors feed.

Kammerer gives a long account of the various conditions to which he subjected both species. The treatment was complicated in many ways, but the essential statements are, as regards S. maculosa, that when no water was provided in which the young might be born, they were dropped on land, larger and in a later stage of development and of a darker colour than is normal; that the larvae so born gradually diminished in number until only two were deposited in each breeding-period; that dissection showed that the other ova degenerated to form a yolk-material. The larvae so produced reached maturity. The summary of results describes their behaviour, stating that they produced:

(a) In water, either (1) very advanced, large-headed larvae 45 mm. long (instead of 25-30 mm.) with gills already reduced, which had awkward, embryo-like movements, and in some few days metamorphosed into small perfect salamanders; or (2) moderately advanced, properly proportioned larvae, 40-41 mm. long, provided with large gills of (at first) intrauterine character, which were reduced during aquatic life.

(b) On land, small (26 mm. long) larvae with rudimentary gills, having the body rounded instead of being flattened from above downwards, and an elongated narrow head, which were unable to live in deep water. These larvae changed to the salamander colour in 10-12 days, and after four weeks metamorphosed into salamanders 29 mm. long.

(c) In the foregoing cases the experimental conditions were not continued, or in other words, basins of water were provided in which they could spawn. But if the experimental conditions are continued, these Salamandra maculosa which were born newt-like (viz., not in a larval condition), are themselves newt-bearing from the first time they give birth, using the dry land, and bringing forth only two young, the normal number for the births of S. atra. These young are 40-41 mm. long, and are dark-coloured, resembling greatly the normal new-born S. atra.

This epitome of the observations illustrating the inheritance of acquired characters has been very widely quoted, and may not unnaturally be taken to summarize a wide experience of the modified animals. Reference to the details given in the same paper shows that, as alleged, each of the four types of behaviour enumerated was witnessed once only in the case of each of four females, no two agreeing with each other. As to the number of the males or their habits nothing is said. The first female, a (1), bore five young; the second, a (2), bore two, of which one was a partial albino; the third, b, produced four young; and the fourth, c, two as already stated.

In the case of c the details show that the female gave birth immediately after being transferred from the open-air terrarium to one indoors, which contained no basin of water. This is the example of the consequences which follow on a continuance of the experimental conditions.[17]

As regards S. atra the converse is reported. Various means were used to induce them to eject their young prematurely in water, such as massaging the sides of the mothers, or raising the temperature to 25° or 30° C., with various degrees of success. But afterwards it was found that specimens collected wild at an elevation of about 1,000 metres responded to much simpler treatment, and gave birth prematurely in water when they were kept in a large shallow basin of water not so deep but that they could everywhere touch the bottom with their feet and keep their heads above the surface. With specimens collected at higher elevations this treatment was inoperative, and the suggestion is made that S. atra at the lower confines of its habitat partakes more of the nature of maculosa than do the individuals from greater heights; for Kammerer argues that pools suitable for breeding must be more uncommon at those elevations than they are lower down.

In the earlier paper[18] Kammerer states that newly caught females of S. atra often give birth in the water, and show an undoubted preference for doing so. He describes also how he once saw several females, wild in their natural habitat, lay their young in a rain-puddle at 1,800 metres elevation, but the larvae thus born were fully formed.

When the deposition of the young as larvae has become "habitual"[19] with S. atra, three to nine larvae may be produced at one spawning period, from 35 to 45 mm. long, with gills at most 8 mm. long, and a tail-fin 2-3 mm. broad. Such larvae are generally coffee-brown, or grey (instead of black), and show other minor differences.

The summary states that when grown to maturity they become in their turn larva-bearing, and go into the water to bring forth. Their young are more than two (3 to 5 being the numbers observed) with a length of 33-40 mm. or of 21-23 mm. at birth. They are light grey, spotted (mottled with lighter and darker colour), have relatively short gills (8 to 9 mm. at most) and a broad tail-fin (3 mm. wide). At metamorphosis they are relatively long (44 mm.) and one of them had some yellow pigment.

Here again this summary is, as a matter of fact, describing the behaviour of two mothers, of which one produced three, and the other five young.

To my mind these experiments suggest that the reproductive habits of both species, if closely observed, will be found to be subject to considerable variation, and I think it not impossible that each species is, especially in confinement, capable of being a good deal deflected from its normal behaviour. Moreover, there seems to me no great improbability in the idea that there is an interdependence between the number of young and the stage of maturity in which they are born. But, at the same time, the case as told by Kammerer strikes me as proving too much. If each species is so sensitive to conditions that the normal procedure is gravely modified in one generation, and if that modification can reappear in a pronounced form in the next generation without a renewal of the disturbing conditions, it becomes extremely difficult to understand how the regularity which each species is believed to display in nature can be maintained. Surely both species might be expected to be in confusion. From a passage in Kammerer's earlier paper (1904, p. 55) on the subject, I infer that he also would expect considerable irregularity in the natural behaviour, but that he has not investigated the point.[20]

3. Modification of the Colour of Salamandra maculosa induced by Change in the Colour of the Soil on which the Animals were kept.—Kammerer speaks of this as the most convincing of all his experiments on the transmission of acquired characters. So far, however, no full account of them has been published.[21] The statement is that when salamanders are kept in yellow surroundings the yellow markings gradually in the course of years increase in amount relatively to the black ground colour. Conversely by keeping the animals on black garden soil, the yellow may be greatly diminished in quantity until it largely disappears. (The account in Natur adds that very moist conditions also favour the increase of yellow, and that with less moist conditions the yellow diminishes.) From each kind, the (induced) yellower and the (induced) blacker, a second generation was raised, on soil of neutral colour, and each family was later divided into two parts, half being put on black and half on yellow ground.

As regards the offspring of those which had lived on black soil no positive result had been reached up to the date of publication, but it is stated that these young resembled their parents in having the yellow distributed in irregular spots.

As regards the offspring of those which had lived on yellow soil the account follows up the story of that part of the offspring which were put on yellow soil again. It is stated that these, though derived from parents with irregular spots, developed the yellow as longitudinal bands.

This account is given with slight differences of expression in the three places to which I have referred. On returning from Vienna in 1910 I consulted Mr. G. A. Boulenger in reference to the subject, and he very kindly showed me the fine series from many localities in the British Museum, and pointed out that in nature the colour-varieties can be grouped into two distinct types, one in which the yellow of the body is irregularly distributed in spots and one in which this yellow is arranged for the most part in two longitudinal bands which may be continuous or interrupted. The spotted form is, as he showed me, an eastern variety, and the striped form belongs to western Europe. Mr. E. G. Boulenger[22] has since published a careful account of the distribution of the two forms. The spotted he regards as the typical form, var. typica, and for the striped he uses the name var. taeniata. The typical form occupies eastern Europe in general, including Austria and Italy, extending as far west as parts of eastern France. The var. taeniata is found all over France, excepting parts of the eastern border, Belgium and western Germany, Spain and Portugal. Of the very large series examined there was only one specimen (Lausanne) which could not with confidence be referred to one or other of the two varieties. Mr. E. G. Boulenger points out that both varieties inhabit very large areas, and live on soils of most different colours and compositions. Both are liable to variations in the amount and the shade of the yellow, but that any suggestion that taeniata belongs especially to yellow soils and typica to black soils is altogether inadmissible. He expresses surprise that Kammerer should not allude to these peculiarities in the geographical distribution of the two forms. He suggests further that it is more likely that some mistake occurred in Kammerer's observations than that the east European typica should, in the course of a generation, have been transformed into the west European taeniata by the influence of yellow clay soil.

In his last paper on the subject Kammerer states incidentally[23] that he has found the striped form recessive to the spotted. No evidence for this statement is given, and I have not found any other reference to crosses effected between the two natural types. If, however, this representation is correct, it is conceivable that the production of taeniata from typica was in fact the re-appearance of a recessive form. The plate which Kammerer gives in illustration of his modified parent figures a single animal at four stages, and though it is certainly more like the spotted than the striped form, it has a certain suggestion of the striped arrangement, such as I can well imagine being produced in the heterozygote.[24]

In continuation[25] of the experiments on the colour of S. maculosa Kammerer publishes an account of elaborate experiments in grafting ovaries of the various forms, modified and unmodified, into each other, and describes the offspring which followed. Before pursuing this part of the inquiry I am disposed to wait until the earlier steps have been made much more secure than they yet are.

More recently Kammerer has published similar statements in regard to the inheritance of characters induced in various lizards by keeping them in abnormal temperatures, high and low. The changes induced affected in some species the colours, in others the reproductive habits. Respecting these examples I feel the same scepticism that I have indicated in regard to the others, somewhat heightened by the fact that insufficient evidence is given both regarding the behaviour of these various species in captivity when not subjected to abnormal temperatures, and in the wild state.

Respecting this part of the evidence Mr. G. A. Boulenger has lately published a criticism[26] from which I extract the following passages. Referring to a previous note[27] on the question of the melanism of the various insular forms of Lacerta muralis he writes: "I also alluded (l. c.) to the theories that have been propounded to explain the melanism of various insular forms. This is a subject which has been lately taken up by Dr. Kammerer at the Biologische Versuchsanstalt in Vienna, and he claims to have produced nigrinos artificially by a very strong elevation of the temperature, accompanied by extreme dryness. Dr. Werner[28] has already opposed his own experiments to those of Kammerer, artificial melanism having been produced by him in Lacerta oxycephala by keeping two very light specimens from Ragusa for a whole summer in very damp conditions. Neither is Kammerer's theory in accordance with the distribution of the black lizards, as pointed out by Werner. Kammerer also finds that those forms which are known to produce melanic races in a state of nature, lend themselves more readily than the others to the success of his experiments. But he shows himself misinformed when he states that the variety called Lacerta fiumana belongs to the category of those of which black forms are not known. He overlooks the fact, first pointed out by Scherer in 1904, and which I can confirm, that the black lizard from Melisello near Lissa in the Adriatic is unquestionably derived from the lizard from Lissa, which he correctly regards as not separable from L. fiumana...."

"Another colour modification which Dr. Kammerer states that he obtained by raising the temperature is the assumption by the female of the typical Lacerta muralis of the bright red colour of the lower parts which often distinguishes the male from the female, and which was not shown by the individuals of the latter sex kept by him under normal conditions. He quotes various authorities to show that the lower parts are never red in the females, but he has omitted to consult others who say the contrary. Thus Bedriaga (1878 and 1879) remarks that a so-called var. rubriventris of the typical wall lizard has the lower parts red in both sexes."[29]

In reading such papers as those of Semon or Kammerer the thought uppermost in my mind is that to multiply illustrations of supposed transmission of acquired characters is of little use until some one example has been thoroughly investigated. If we had certain assurance that even a single unimpeachable case could be repeated at will, the whole matter would assume a more serious aspect. If, for instance, Kammerer were able to show us Alytes males with horny pads on their hands, it would be something tangible; still more, if the experiment were repeated by others until no doubt remained that the offspring of Alytes which had bred in water for some three generations did acquire these pads and that they could transmit these novelties to descendants raised in normal conditions. Till evidence of this kind is published by at least two independent observers investigating similar material, I find it easier to believe that mistakes of observation or of interpretation have been made than that any genuine transmission of acquired characters has been witnessed.

Meanwhile there is no denying that the origin of adaptational features is a very grave difficulty. With the lapse of time since evolutionary conceptions have become a universal subject of study that difficulty has, so far as I see, been in nowise diminished. But I find nothing in the evidence recently put forward which justifies departure from the agnostic position which most of us have felt obliged to assume.[30]

Appendix to Chapter IX.

Professor G. Klebs, as is well known to students of evolutionary phenomena, has for several years been engaged in investigations relating to the inheritance of acquired characters. In his many publications on the subject the issue has always been represented as more or less uncertain.

Desiring to know how the matter now stands according to Professor Klebs' present judgment I wrote to him asking him to favour me with a brief general statement. This he most kindly sent in a letter dated 8th July, 1912.

As such a statement will be read with the greatest interest by all who are watching the progress of these studies I obtained permission to publish it as follows:

8. Juli 1912
Ihre liebenswurdige Anfrage will ich sehr gern beantworten, obwohl ich sie nicht so beantworten kann wie ich erwÜnschte. Ihr Skepticismus in der Frage der Uebertragung erworbener Charactere auf die Nachkommen ist nur zu berechtigt. Meine Versuche mit Veronica sind nicht beweisend, da es mir bisher nicht gelungen ist eine einigermasse konstante VarietÄt mit verlaubten Inflorescenze zu erzeugen. In Bezug auf mein Semper vivum bin ich allerdings noch heute der Meinung dass die starke kÜnstliche VerÄnderung der BlÜte einen Einfluss auf einzelnen Nachkommen gehabt hat. Ich habe seither nichts darÜber verÖffentlicht: die Mehrzahl der anormalen gefÜllten BlÜten war leider steril. Von einem weniger verÄnderten Exemplar erhielt ich einige SÄmlinge, aber sie haben noch nicht geblÜht. Es kann sich in diesem Falle nur um eine Nachwirkung in der ersten Generation handeln, vergleichbar jenen FÄllen in denen Samen von BÄumen aus den hohen Alpen in der Ebene gewisse Nachwirkungen zeigen. Aber es ist bisher kein sicherer. Fall bekannt in den der kunstliche herbeigefÜhrte Charakter mehrere Generationen hindurch unter der gewÖhnlichen "normalen" Bedingungen Übertragen worden ist.

Auf der andere Seite sind diese negativen Resultaten nicht entscheidend. Denn wie wenig ist in dieser Beziehung Überhaupt ernstlich versucht worden! Und zweifellos geht die Sache nicht so einfach.

Ich versuche es mit anderen Pflanzen weil ich der Meinung bin dass es mÖglich sein mÜsse wenigstens solche neuen VarietÄten zu erzeugen, wie sie die GartenvarietÄten entsprechen.

Aber bis jetzt leider sind die Versuche nicht gelungen, weder mir noch irgend einem anderen.